Process optimization in batch scheduling and distillation column sequencing

Process optimization plays an important role in the chemical process industry. In its most general form, process optimization is a multiobjective and multivariable optimization problem of an extremely complex nature. These complex optimization problems are divided into smaller problems that are easier to solve using heuristic techniques or mathematical optimization methods.; This thesis is concerned with two optimization problems of great importance to the chemical industry, batch scheduling and distillation column sequencing. Simulated annealing is used as the optimization technique to obtain solutions by framing the problems as single objective, multivariable optimization problems.; In batch scheduling, two distinct problems are considered. The first problem involves the scheduling of multiple products on a network of single-stage, unrelated parallel units. The performance criterion used is tardiness minimization. The solution methodology incorporates sequence dependent clean-up times, varying processing rates for different products on different units and product to unit assignment constraints indicating a set of feasible processing units. The second problem considers cyclic multiproduct sequencing on continuous parallel lines. The objective function is based on processing cost, sequence dependent transition times and costs, inventory cost and a profit boost associated with creating idle times on the continuous processing lines so as to incorporate fluctuating market requirements. The solution methodology for this problem also incorporates varying processing rates for different products on different units and constraints involving product to unit assignments indicating a set of feasible processing units.; The problem of distillation column sequencing involves the separation of a multicomponent feed stream into several multicomponent product streams. It is recognized that a cost efficient method will introduce blending of certain multicomponent bypass streams to reduce the separation mass load. The approach incorporates stream splitting, mixing, and bypassing of multicomponent streams along with the use of multiple columns performing similar separation tasks. The algorithm uses a general structure for the sequence of distillation columns, sets upper bounds on flowrates within a feasible search space, and is implemented in a computationally efficient manner.; This work also compares two optimization algorithms, simulated annealing and threshold accepting by examining their convergence properties using time inhomogeneous Markov chains.